The National Health Interview Survey (NHIS) is an important source of health information about the U.S. population. Public-use microdata covering over 60 years of the NHIS are available. Typically included in public releases are several sequentially defined survey weights, e.g., base, non-response adjusted and final calibrated, and for variance estimation pseudo strata and PSUs (masked to avoid geographical disclosure). The suggested design-based variance estimation procedure is to treat the final multiple-adjusted weight as a base sampling weight and to use linearization methods along with the strata-PSU “S^2- forms” provided in standard survey software packages. While easy to implement, this variance estimation method ignores the fact that the final provided weight is a function of the sample and thus subject to variability. In this paper we demonstrate public-use methodology that a data user can apply in creating non-response, subsampling and calibration adjustments for survey weighting along with replicate weighting techniques, e.g., bootstrap, jackknife, and BRR. Using such methods may better reflect the sampling variability over a wide range of survey estimators.